An original approach for gas chromatography-olfactometry detection frequency analysis: Application to gin
Résumé
Gas Chromatography-Olfactometry (GC-O) is a technique that lies halfway between physicochemical and sensory analysis and involves the perception of volatile flavour compounds, separated by gas chromatography, by the human nose. Of the different GC-O procedures available, detection frequency has been proved to be more rapid and more repeatable, even with an untrained panel. This characteristic regarding the panel is often not considered when dealing with the sensory attributes determined by assessors. An original approach to GC-O using the detection frequency procedure has been developed and tested on two types of gin and made it possible to benefit from sensory data. The panel consisted in 10 assessors (five male and five female) with prior GC-O experience. They generated odour events each time they perceived an odour. In addition, they were asked to provide as precise a description as possible of the smell they perceived. The 2589 odour events thus generated were then grouped into 209 odorant areas using a specially-developed software, based on the closeness of their linear retention indices. These odorant areas were then associated with compounds identified using Gas Chromatography Mass-Spectrometry. The intensity of a compound was defined as the number of odour events in the corresponding odorant area. Thirty height compounds, mostly terpenoids, were identified as having an impact on gin aroma and enabled differentiation of the two products. Among these 38 compounds, those belonging to the two products were described as constituting a "gin base" , while the others were described as "typical compounds" . This procedure also enabled the identification of compounds whose perceived intensity was highly sensitive to changes in concentration. Finally, an odour classification wheel was developed to categorise the descriptors obtained in 12 supercategories. An odour profile was generated for each odorant area, based on the occurrence of a descriptor in a supercategory. These odour profiles were either homogeneous (with one dominant odour) or heterogeneous (with no dominant odour). This approach enabled a quantitative sensory description of the odorant areas generated by GC-O.